Table 1 Summary of parameters for neutral atom quantum computers, representing the current state-of-the-art with an extrapolation to the future given in brackets. The values reported here reflect the status of the devices run by neutral atom quantum computing companies, with the perspective of being used for applications already now or in the near future. A description of each parameter can be found in Sect. 3.1

Amount

∼100 [20–23] (∼1000 for 2024 [44, 45])

Connectivity

10:1–20:1 [20], (50:1–100:1 possible in principle)

Multiple states (i.e., qudit)

In principle possible

Trap lifetime

10–60 s [22, 23, 29, 30] (up to 6000 s with cryostat [39])

Decoherence times (electronic spin)

\(T_{1} \sim 4\text{ s}\), \(T_{2} \sim 1\text{ s}\), \(T_{2}^{*} \sim 4\text{ ms}\) [21, 22]

Decoherence times (nuclear spin)

\(T_{1} \gg 5\text{ s}\), \(T_{2} \sim 40\text{ s}\), \(T_{2}^{*} \gg 3\text{ s}\) [23]

List of gates

Single-qubit rotations, CZ → CNOT [20, 38], SWAP, CPHASE [37]

>2-qubit gates:

CCZ → Toffoli /CCNOT [20, 38], \({C}_{k}{Z}\), generalisation to k control qubits [37]

Parallelism

Apply the same gate on multiple Qubits simultaneously: Single-qubit rotations, CZ [21] (Multiple gates on multiple Qubits)

1-qubit gate

0.996–0.999 [21, 22]

2-qubit gate

0.955–0.995 [22, 38, 42]

Readout

≳0.95 [46]

Preparation

Trap occupation probability (after rearrangement): 0.988 [47]

Success probability for defect-free array ∼0.75, depending on size of array [47]

1-qubit gate

∼2 μs (π-pulse) [21]

2-qubit gate

∼400 ns (CZ) [21, 38]

Preparation (incl. rearrangement)

∼400 ms [47]

Readout

∼10 ms for fluorescence imaging [46] (∼6 μs using a collective readout scheme) [46]

Required infrastructure

Vacuum cell and pumps, lasers, optical elements, microwave sources, signal generators and modulators, magnetic field coils. Cooling the setup with a cryostat can improve vacuum quality and increase trap lifetimes [39]. (Rack-level implementation possible [48])

Calibration

Rearrangement at beginning, no calibration of individual qubits necessary

Specificity

Shuttling operations [21]

Access

Via the cloud [49, 50] and on-premise

Quantum computing paradigm

Gate-based (digital) quantum computing, digital-analogue quantum computing, quantum annealing, analogue quantum simulation